Russ: Welcome back to The BusinessMakers Show. My guest now is Reid Smith, Co-founder and CEO of i2k Connect. Reid, welcome to The BusinessMakers Show.
Reid: Thank you, Russ. It’s a pleasure to be here.
Russ: You bet. Tell us about i2k Connect.
Reid: Well, we started this company a couple of years ago to try to deal with the trouble people have finding, analyzing, and extracting the vital information from what are called unstructured documents; the kind of thing that you use every day with Microsoft Office products or with PDF documents. We found that people are trying to do this by hand. No one wants to do it. It’s a task that librarians used to take on, but no longer (Russ: Librarians). Now, you and I, and all of our colleagues are faced with being our own librarians, and we don’t do a great job at it.
Russ: Well, I certainly underscore that. Not only in the amount of documents I deal with today, but I have a business history, did some data document discovery in the legal world. Optical character recognition, back when we had that, and had armies of people building databases, but I assume you don’t do it with armies of database builders.
Reid: Few people, perhaps, aside from your former company, have a revenue stream that would support armies of people to do that by hand. So, yes, the things that have really changed over the years, I mean, first of all, massive changes in the amount of computing power that’s available. Second, the web; massive amounts of content that can be mined to apply the latest in artificial intelligence techniques, machine learning techniques, which everyone is aware of today. But also, and what’s called natural language; technology that can read documents, not to the professional level that you might, but to a skimming level that’s good enough for some of our purposes.
Russ: Ok, and speaking of your purposes, who are your major customers or industries?
Reid: Well, the industry we focus on is oil and gas. I was hesitant to give you this number, but between me and the rest of the team, we’ve got over 100 years of oil and gas experience.
Russ: Well, congratulations, but it’s a large team so,
Reid: Not as large as you might think. No, and so we started by focusing on oil and gas, upstream and downstream; we’ve had experience in both, but we also found, or others found us, in an area that’s interesting, in supply chain risk management. And then, an area that we didn’t expect at all was in healthcare and wellness. Just think of it as the consumer side of healthcare.
Russ: Ok, so all of those can be document intensive industries. In healthcare there is this movement to digitize medical records, but you’re talking about still dealing with the original documents and paper documents?
Reid: Well, interestingly enough, we’ve been working with a small company in San Diego. The name of the company is Tovifit, and what they do is to supply content to support employee wellness programs. So you can imagine that your, the HR department for your company has decided that it’s very beneficial, of course, for the employees to have better health than they do today (Russ: Absolutely.), and one of the ways to help them get to that state is through education. And so, this company writes its own content and, what we’ve started to do with them is mine the content that’s available on the internet.
Russ: Ok, so they’re giving to you Word documents and that sort of thing.
Reid: What we’re doing is subscribing to a lot of internet news feeds, and we’re figuring out which of those feeds are important for healthcare, and then the particular aspects of healthcare. Is this something that you should read from the point of view of weight management, or stress management, or forms of relaxation; yoga, for example. So, our system is able to classify the content at a fairly detailed level. And so, it can serve up to you content that is quite specific to where you are in the sort of fitness spectrum, and the things that interest you, for instance. And we do that same thing for supply chain risk management and in oil and gas.
Russ: Ok, wow. So, like back in the healthcare, this particular customer, what percentage of the data that you’re dealing with is data that you went out and retrieved for their purposes?
Reid: Oh, that’s a very good question; almost all of it. I’m not sure I could give you a percentage, but it’s a small company, so we don’t have vast resources to be authoring content, but they are doing some, and that was their original model. And what we’ve done is augment the model with content that is available, that does not require them to scale up, resource wise, quite as quickly as they might otherwise need to do.
Russ: Ok, so you’re helping them have something for their clients, patients, and customers to use; interesting. So, let’s move back to oil and gas then. What kind of specific applications are you applying your technology to in oil and gas?
Reid: Well, one of them is to help oil and gas companies find the documents they need to answer particular questions. And one scenario is, when any kind of job that is done, there will be a variety of data capture. Some of the data, or so called structured data, they appear in databases and apps, and the industry has years of experience in dealing with that and managing that, but you know some of the other data is actually found in PowerPoint files, and PDF files, and text files. We don’t do such a good job at that, and for a couple of reasons. One of them is, the apps available to us are the standard word processing apps, or Excel, or something like that, and none of those files have been marked up to provide all the metadata that a search engine needs to help you find the data. It’s quite different from what you see on the internet.
On the internet, people like you and like me are working very hard to ensure that our content is findable by anyone, but that is not the case inside of companies. We’re depending on individual engineers, geoscientists, HR professionals to do that work; oh, by the way, in addition to all of the other things that they do. And they just don’t have the time to do it, let alone be trained to do it.
Russ: So how do you do this if you say you don’t do PDF, PowerPoints, and Word docs very well. How does that information become accessible?
Reid: So, in the public domain, or in the industry domain, it is accessible to those who subscribe. And, of course, inside your company, the files are accessible. The issue is mining the files to add the metadata that a search engine requires, or an analytics tool requires to make sense of them. And that is exactly the kind of work that we do.
Russ: Ok, very interesting. And how old is the company?
Reid: Well, we started at the end of 2013, but we started actually much earlier in the sense that three of us had done some pro bono work for one of the major professional societies. In fact, the Premier AI, Artificial Intelligence Society, and what we had started to do was to deliver a new service. Originally the new service was done by hand. A person would go and surf the internet, classify documents, and say, these are interesting. And it’s about this particular kind of AI, its machine learning, or its natural language, or its robots. Well, that person, after a while, retired, and the question for us was, do we shut it down or do we automate it? Well, we automated it, and we have now for the last five years been running an automated news service.
It’s kind of interesting because its artificial intelligence used for an artificial intelligence news service (Russ: That’s pretty cool.). And so we thought that, armed with experience I had working with major service company and a medium sized oil and gas company led us to believe that maybe there were other applications of this technology.
Russ: Well, it seems to me this picture that’s being painted in my mind, after understanding your business a little bit better, is that we are just inundated these days with information. So, you not only go out and get the pertinent information, but do you kind of weed out that information that’s not pertinent?
Reid: Absolutely, and so we subscribe to hundreds of news feeds. To take the internet case just to start with, and on any given day you will find thousands of news stories in the sources that you may personally be watching. Or, you’re watching them through Google News or something. An example that is maybe interesting: on a particular day, our system pulled in 2300 news stories. Over 100 were from oil and gas sources. So the kind of sources you, you know, with your job, perhaps as a production engineer or a reservoir engineer, would be expected to be monitoring.
On that given day though, only one of them was about hydraulic fracturing in the Permian Basin (Russ: Which might be my area of interest) and it turns out—exactly, and our story, that is your area of interest. So the tuning, the filtering, the extraction of the essentials from the clutter you got is something like 1 over 2300, you know, or 1 over 100. It’s a big factor. We’re giving you just what it is that you need to do your job so that you can get only the data required to recognize problems and to address them (Russ: In my area.), and not have to go and search through those things by hand to find them.
Russ: Wow, it seems like it’s almost a service that might be of interest to individuals at some point in the not too distant future.
Reid: It’s an interesting question, and of course that thought crosses our minds. The difficulty there is just the breadth of knowledge you have to have to serve the consumer market. Now, as it has turned out, we have had a little about a lot of areas in order to do even what we’re doing properly, because the news sources that people are looking at, they cover everything from sports to entertainment to, oh, by the way, fracturing in the Permian Basin sometimes. And so what we learned is that our original model of, get very good in a very narrow domain; that has not worked out quite the way we had expected. We’ve had, yes we do that, but we’ve also had to learn a little bit about a lot of areas in order to do this job properly.
Russ: Right. Well, Reid I really appreciate you sharing your story with us.
Reid: Well, thank you, Russ. It’s great to spend time with you.
Russ: You bet. And that wraps up my discussion with Reid Smith, Co-founder and CEO of i2k Connect, and this is The BusinessMakers Show.
brought to you by